The goal of this study is to establish which immuno-suppressive therapy - methotrexate or mycophenolate mofetil - is more effective as a first-line, corticosteroid-sparing agent for the treatment of non-infectious uveitis in a masked, randomized trial. Uveitis is a significant cause of vision loss and morbidity in the United States and the word. Due to the host of local and systemic side effects associated with long-term corticosteroid therapy, other immunosuppressive therapies are frequently used as corticosteroid-sparing agents in patients who are unable to successfully control inflammation with a reduction in corticosteroids. These include the antimetabolites, calcineurin inhibitors, alkylating agents, and biologic drugs. Cost and morbidity associated with uncontrolled inflammation make the selection of an effective initial corticosteroid- sparing agent extremely important. When corticosteroid-sparing therapy is required, two of the most commonly used first-line immunosuppressive agents are the antimetabolites methotrexate and mycophenolate mofetil; other classes of drugs such as biologics are typically reserved for refractory cases in which other treatments have failed because of their cost and potentially dangerous long-term side effects. It is common practice for patients requiring a steroid-sparing agent to be treated first with methotrexate and then switched to mycophenolate mofetil in the event of treatment failure. However, results from retrospective case series indicate that uveitis patients may be much more likely to achieve controlled inflammation and tolerate treatment with mycophenolate mofetil. There have been no prospective randomized, controlled trials to systematically determine which antimetabolite is more clinically effective as initial steroid-sparing therapy, and there is no evidence basis for deciding whether to switch to another antimetabolite or to move on to another class of immunosuppressives if patients fail one antimetabolite. In this multicenter, masked, randomized trial, we will evaluate the safety and efficacy of methotrexate and mycophenolate mofetil as initial corticosteroid-sparing therapy for patients with non-infectious uveitis, as well as evaluate the clinical effectiveness of switching agents aftr initial treatment failure. Patients will be enrolled at UCSF, OHSU, and Aravind Eye Hospital.